A once-per-day, drug-in-food protocol for prolonged administration of antiepileptic drugs in animal models


Address correspondence to F. Edward Dudek, Department of Physiology, University of Utah School of Medicine, 420 Chipeta Way, Suite 1700, Salt Lake City, UT 84108, U.S.A. E-mail: ed.dudek@hsc.utah.edu


Purpose:  Convenient and effective methods for administering potential antiepileptic drugs (AEDs) chronically should facilitate many experiments in animal models of chronic epilepsy with spontaneous recurrent seizures. This proof-of-principle study aimed to optimize a once-per-day, drug-in-food protocol by testing the effect of carbamazepine (CBZ) on the frequency of convulsive seizures in rats with kainate-induced epilepsy.

Methods:  Adult male rats were given repeated low-dose kainate injections until convulsive status epilepticus persisted for >3 h. After the rats developed spontaneous recurrent seizures, food pellets with CBZ (30, 100, or 300 mg/kg/day) were provided once per day in three 2-week trials (n = 7–9 rats) involving 5 days of CBZ or control treatment, separated by two recovery days within a trial. The total amount of food provided and consumed per day corresponded to a normal caloric diet (60 g/kg/day).

Key Findings:  When provided once per day, all animals ate the CBZ-containing food irregularly but continuously throughout the 24-h day. With this daily feeding protocol, CBZ significantly reduced the frequency of spontaneous convulsive seizures in a dose-dependent manner. It is important to note that the effect of CBZ was consistent across the 5 days and throughout each day of the trials. With food administered at 9:00 a.m., and blood assayed at 5:00 p.m., higher food levels of CBZ resulted in higher plasma concentrations of CBZ.

Significance:  This AED-in-food protocol is simple, efficient, inexpensive, reliable, and noninvasive; it allows easier long-term drug administration and is less stressful and more humane than other methods of AED administration.

Animal models of epilepsy with spontaneous recurrent seizures may be useful for discovery of new antiepileptic drugs (AEDs). Improved methods of AED administration that minimize stress during treatment and still allow long-term maintenance of adequate plasma levels are needed (Loscher, 2007). Conventional methods include intraperitoneal injections and oral gavage, which are stressful, and thus may increase seizure frequency, particularly when drugs must be administered multiple times per day (Loscher, 2007). Poor taste and insolubility in water are potential problems with mixing drugs in drinking water (Loscher & Schmidt, 1988). Other approaches include drug infusion through a gastric tube (Jones et al., 2009), subcutaneously implanted osmotic minipump (Glien et al., 2002), and an implanted catheter with an external pump (Bertram et al., 2005; Loscher, 2007). However, a simple, efficient, inexpensive, reliable, and noninvasive procedure for AED administration would be useful in epilepsy research.

Carbamazepine (CBZ) is arguably the most widely used AED; it is clinically effective against partial and generalized convulsive seizures, but its aqueous insolubility and short half-life are major drawbacks for designing chronic experiments that require prolonged administration (Loscher & Honack, 1997). Therefore, CBZ is an appropriate “test AED” for a “proof-of-principle” study on chronic AED administration in animal models of epilepsy.

In a previous study (Grabenstatter et al., 2007), CBZ was mixed in chocolate-flavored food pellets and single food administrations resulted in comparable anticonvulsant efficacy as single intraperitoneal administrations (with equal doses). When CBZ was administered in food at three, equally spaced times per day, spontaneous recurrent convulsive seizures were strongly suppressed. However, a three-times-per-day feeding protocol is labor-intensive; therefore, we aimed to optimize a once-per-day protocol. Because the study of Grabenstatter et al. (2007) tested only a high-dose of CBZ (i.e., 300 mg/kg/day), it was also important to determine if a once-per-day protocol would be effective at lower doses, and to determine if the drug effect on spontaneous seizures would persist throughout the day. These experiments strongly suggest that a once-per-day, drug-in-food protocol using specially prepared flavored food would be useful for chronic testing of new AEDs and for other experiments that require long-term AED treatment.


Kainate treatment and induction of status epilepticus

All procedures with animals were approved by the University of Utah Animal Care and Use Committee. The animals were subjected to the repeated, low-dose protocol of kainate-induced status epilepticus (see Hellier & Dudek, 2005; Dudek et al., 2006; and see Data S1).

Selection of experimental animals

Beginning 1 week after kainate-induced status epilepticus, the animals were monitored 6 h per week for the next 4–6 months to identify animals that exhibited spontaneous recurrent convulsive seizures. All animals that exhibited spontaneous convulsive seizures were then continuously video-monitored for periods of 1 week for final selection for the AED trials, and all rats used in this study were confirmed to have had frequent (i.e., 5–14 seizures per day) class III–V convulsive seizures as a prerequisite for inclusion in the study. Two animals died after the CBZ 300 mg/kg/day trial and before the 100 mg/kg/day trial, and one animal stopped eating and drinking, so it was not used in the study; therefore, these three animals were replaced for the CBZ 30 and 100 mg/kg/day trials.

CBZ-formulated food and the repeated-measures, cross-over protocol

The amount of food required for maintenance of body weight was 60 g food/kg/day (Grabenstatter et al., 2007). CBZ was formulated in food pellets as 5-mg CBZ per 1-g food pellet (Bio-Serv, Chocolate Mini-Treats, Catalog # F05472; Frenchtown, NJ, U.S.A.). Control pellets were identical, but without CBZ. Based on the dose, rats were provided with the appropriate number of CBZ-containing and/or control pellets (e.g., 20 CBZ pellets and 40 control pellets for 100 mg/kg/day), and they ate all of the pellets. Food pellets with CBZ (30, 100, or 300 mg/kg/day) were provided in three 2-week trials (n = 7–9 rats) involving 5 days of CBZ or control treatment, each followed by two recovery days (Fig. 1A; also see Fig. S1). CBZ-containing and/or control food was provided in a single feeding each day (9 a.m.), which was 3 h after the start of the light cycle. The animals were weighed every fourth day to determine any possible effect of CBZ-containing food versus control pellets on body weight. No difference in the progressive increase in body weight was seen during CBZ and control treatment.

Figure 1.

(A) Time line used for CBZ-in-food trial. Time line of the trials for different CBZ doses after kainate-induced status epilepticus. The animals were allowed to develop spontaneous recurrent seizures over a period of 4–6 months after kainate-induced status epilepticus. Each CBZ trial required 2 weeks, and was separated by 2 weeks. (B) Feeding pattern of rats with kainate-induced epilepsy when fed with CBZ-containing or control food. The data are presented as mean number of eating events during 2-h epochs (e.g., 9–11 a.m.) across 24 h. Vertical bars, standard error of the mean (SEM).

Continuous monitoring of spontaneous recurrent convulsive seizures

Every animal was monitored continuously throughout the protocol on three 8-h DVDs [(1) 8 a.m.–4 p.m., (2) 4 p.m.–12 a.m., and (3) 12 a.m.–8 a.m.] using an infrared camera system (EZWatch Pro, Louisville, KY, U.S.A.). A trained technician blinded to the treatments and dates viewed the animals’ behavior. The DVDs were observed in the fast-forward mode for any activity suggestive of a seizure (running, jumping, rearing, lordosis, erect tail, and so on). If any seizure-like activity was seen, the DVD was stopped, reversed, and watched in real time to evaluate the behavior for possible seizures.

Estimation of CBZ plasma levels

Separate animals (n = 15, not treated with kainate) were used for the studies aimed at measuring plasma levels of CBZ. The 15 animals were divided into three groups (30, 100, and 300 mg/kg), and 15 samples were obtained for each of the three doses (45 samples total). For estimation of CBZ in plasma, blood was collected at 5 p.m. (8 h after the time of food administration) from the lateral saphenous vein after brief anesthesia. CBZ was estimated in plasma using reverse-phase chromatography. Briefly, rat plasma (25 μl) was diluted 1:1 with methanol containing nitrazepam as an internal standard, and extracted with 700 μl of methylene chloride. The methylene chloride layer was evaporated to dryness, and the residue was dissolved in a small volume of methanol. The methanol solution (50 μl) was subjected to reverse-phase chromatography (Supelco Discovery C18; 25 cm × 4.6 mm) at 35°C with 20 min gradient (0–90% acetonitrile) elution and 0.1% aqueous trifluoroacetic acid at a flow rate of 1 ml/min. CBZ (retention time: 15.95 min) was detected by its absorbance at 250 nm, and quantified by area, by comparison with standard curves.

Statistical analyses

Spontaneous recurrent seizures that developed after kainate-induced status epilepticus exhibited heterogeneous variance and followed a non-Gaussian distribution irrespective of control or CBZ treatment. Therefore, a Wilcoxon matched-pairs signed-rank (nonparametric) test was used to analyze the data. A chi-square test was used to analyze the data on the percentage of animals exhibiting spontaneous seizures.


Diurnal pattern of food consumption

Epileptic rats ate intermittently (i.e., irregularly) but continuously throughout the 24-h day when the daily dose of control or CBZ-supplemented food was provided at 9 a.m. An eating event or feeding episode was recorded when animals were observed to eat the pellets at a particular time, as observed in the video recording (see Fig. S2). The mean number of eating events was calculated for all animals during 2-h epochs across 24 h. The number of pellets consumed per eating event was not recorded. The feeding episodes appeared irregularly distributed, but the animals ate throughout the 24-h day, and the presence of CBZ in the food did not appear to affect the timing of the eating episodes. The most eating events occurred in the 2-h epoch beginning 1 h after the beginning of the dark period (7–9 p.m.), and the fewest events were observed in the 2-h epoch beginning 1 h after the beginning of the light period (7–9 a.m.) (Fig. 1B). Therefore, the animals appeared to eat throughout the 24-h day, but light-dark transitions also appeared to affect feeding activity.

Effect of CBZ on seizure frequency over time

The effect of 30, 100, and 300 mg CBZ/kg/day on the relative frequency of convulsive seizures was determined (see the experimental protocol shown in Fig. S1). Relative seizure frequency was calculated as the seizure frequency during administration of CBZ normalized to the frequency of seizures during the corresponding control conditions. For each dose, the relative seizure frequency was calculated from the raw data on control and CBZ treatment (Figs 2 and S3). Overall, convulsive seizure frequency was reduced to about 50% for a CBZ dose of 30 mg/kg/day, and the reduced seizure frequency was statistically significant only on days 2 and 3 of treatment (Figs 2A and S3A). CBZ at 100 mg/kg/day blocked most spontaneous convulsive seizures (Figs 2B and S3B), and CBZ at 300 mg/kg/day completely blocked all spontaneous convulsive seizures throughout the 5 days of treatment (Figs 2C and S3C).

Figure 2.

Oral administration of CBZ, studied at three doses, reduced convulsive seizure frequency. Seizure frequency for 5 days of CBZ-in-food treatment was compared to control food treatment at three doses (AC) using a nonparametric test (i.e., Wilcoxon signed-rank test). The relative seizure frequency (i.e., CBZ normalized to control food) is shown for each dose. (A) CBZ at 30 mg/kg/day reduced the frequency of convulsive seizures by about 50%. (B) CBZ at 100 mg/kg/day blocked nearly all convulsive seizures. (C) CBZ at 300 mg/kg/day completely blocked all convulsive seizures. Dashed horizontal line represents baseline (i.e., no effect). Significant differences between CBZ and control are indicated by asterisks (*p < 0.05, **p < 0.01). Vertical bars, SEM.

Dose-dependent effect of CBZ

CBZ treatment caused a dose-dependent reduction of spontaneous convulsive seizures (Figs. 3 and S4A); the highest dose (300 mg/kg/day) exhibited complete blockade of convulsive seizures. If “responders” were defined as animals whose seizure frequency decreased by ≥50%, 57% were responders with 30 mg CBZ/kg/day, whereas 100% of the animals were responders in the 100 and 300 mg/kg/day groups. All rats (7/7) treated with CBZ at 30 mg/kg/day had some convulsive seizures, whereas only 25% (2/8) of the rats exhibited convulsive seizures during the 5 days of treatment with 100 mg CBZ/kg/day. However, none of the rats (0/8) had convulsive seizures with 300 mg CBZ/kg/day (Fig. S4B). Therefore, the effect of CBZ on convulsive seizures was dose-dependent, when using the once-per-day, drug-in-food protocol.

Figure 3.

CBZ administered in food reduced seizure frequency in a dose-dependent manner. CBZ-in-food treatment for 5 days was compared to control food treatment using a nonparametric test (as above). The relative seizure frequency during drug treatment is shown for each of the three doses. Significant differences between CBZ and control are indicated by asterisks (**p < 0.01, ***p < 0.001). Vertical bars, SEM.

Effect of CBZ on the frequency of convulsive seizures as a function of the light–dark cycle

The anticonvulsant effect of CBZ was assessed as a function of the light–dark cycle (light: 6 a.m.–6 p.m.; dark: 6 p.m.–6 a.m.). CBZ at 30 mg/kg/day did not reduce the relative frequency of convulsive seizures between 6 a.m. and 10 a.m., although it significantly reduced the relative frequency of convulsive seizures during the remaining parts of the light and dark cycle (Figs. 4A and S5A). The lack of a statistically significant effect of CBZ, 30 mg/kg/day, between 6 a.m. and 10 a.m. may have been due to a lack of CBZ-containing food pellets (i.e., the rats may have eaten all of the food before the next 9 a.m. feeding). The anticonvulsant effect of CBZ when administered in this once-per-day, drug-in-food protocol at 100 and 300 mg/kg/day, however, lasted for the entire 24-h period between food administrations, and convulsive seizures were blocked in a similar manner during the light and dark phases of the cycle (Figs 4B,C and S5B,C).

Figure 4.

CBZ-in-food treatment reduced relative seizure frequency throughout the 24-h day. No significant difference in relative seizure frequency was observed between day and night during CBZ treatment (Wilcoxon signed-rank test). The relative seizure frequency (i.e., normalized) for the two foods (CBZ and control) is shown for each dose. The data are presented as the mean of the relative seizure frequency during 4-h epochs (e.g., 6–10 a.m.) across 24 h (A: 30 mg/kg/day, B: 100 mg/kg/day, C: 300 mg/kg/day). Dashed horizontal line represents baseline (i.e., no effect). Significant differences between CBZ and control are indicated by asterisks (*p < 0.05, **p < 0.01). Vertical bars, SEM.

Plasma CBZ levels

The mean plasma levels of CBZ 8 h (i.e., 5:00 p.m.) after the daily feedings at 9:00 a.m. are shown in Fig. 5A. CBZ treatment at 300 mg/kg/day resulted in a mean plasma concentration significantly higher than the 30 and 100 mg/kg/day treatments. CBZ treatment at 30 and 100 mg/kg/day resulted in mean plasma concentrations slightly below the suggested human therapeutic range of 4–10 μg/ml, whereas plasma levels of CBZ were within the therapeutic range when administered at 300 mg/kg/day. The relative frequency of convulsive seizures was not a direct function of plasma levels (Fig. 5B), but the percentage of animals with complete cessation of convulsive seizures did appear to be better correlated with plasma levels of CBZ (Fig. S6). The apparent lack of significant difference in plasma concentration for 30 versus 100 mg/kg could reflect limitations on sensitivity of the assay at lower doses. The relatively small difference in seizure frequency between 100 and 300 mg/kg could reflect the high efficacy of 100 mg/kg on the convulsive seizures (i.e., a saturation effect at 100 mg/kg).

Figure 5.

CBZ-in-food treatment resulted in dose-dependent CBZ plasma levels. (A) Blood was collected 8 h after presentation of the CBZ-containing food. N = 15 per dose. Vertical bars, SEM. (B) A nonlinear association existed between relative seizure frequency and plasma CBZ levels (log concentrations) at different doses (30, 100, and 300 mg/kg/day). Statistical significance denoted by asterisks (***p < 0.001, CBZ 300 mg/kg/day vs. 30 and 100 mg/kg/day; Wilcoxon signed-rank test). Horizontal bars, SEM. Dashed horizontal line represents baseline (i.e., no effect).


The present study significantly extends a report from Grabenstatter et al. (2007) by showing the following: (1) when CBZ-containing or control food was provided once per day in the morning, rats with kainate-induced epilepsy showed an irregular eating pattern, but ate throughout the 24-h day; (2) CBZ administration by this method reduced the frequency of spontaneous convulsive seizures in a dose-related manner, and this effect was evident throughout the entire 24-h period between the feedings; and (3) 8 h after providing the CBZ-containing food, plasma CBZ concentrations were correlated with the drug dose. These data indicate that (1) rats with kainate-induced epilepsy eat irregularly but many times during the 24-h day; (2) CBZ suppressed seizures throughout the 24-h day; and (3) plasma levels of CBZ were dose-related. Although unlikely, the effectiveness of this treatment strategy could be model-specific. This protocol for chronic administration of AEDs in the food is simple, inexpensive, reliable, and noninvasive.

Problems with different methods of AED administration

The once-per-day, drug-in-food method of AED administration reported here offers the obvious advantage of minimizing the handling of epileptic rats, and avoids the stress and pain associated with gavage or intraperitoneal injections, which in turn may provoke seizures (Joëls, 2009). This method should be useful for drugs with short half-lives, which require multiple administrations per day. Because the drug is mixed in chocolate pellets that are readily eaten, it reduces or eliminates problems with drug palatability and solubility. Finally, the drug-in-food method is arguably more humane than other available methods.

Diurnal eating pattern of epileptic rats

Rats normally consume most of their food at night (Fitzsimons & Le Magnen, 1969; Zucker, 1971; Bealer & Johnson, 1980; Loscher & Schmidt, 1988). In the present study, the epileptic rats ate throughout the 24-h day, although the pattern was irregular and a maximum was observed at 7–9 p.m. However, we monitored only the number of apparent eating events and not the amount of food consumed at different times, so the rats with kainate-induced epilepsy may have eaten substantially more food at night than during the day. Seizures in epileptic rats also disrupt the circadian rhythm (Raol & Meti, 1998; Quigg et al., 2001; Bastlund et al., 2005) due to neuronal loss in regions that control the circadian rhythms (Bastlund et al., 2005). A loss of the circadian rhythm could result in a more equal distribution of food consumption between light and dark phases (de Castro et al., 1978; Bealer & Johnson, 1980).

Effect of CBZ on spontaneous convulsive seizures

Previous results on the effect of CBZ in animal models with spontaneous recurrent seizures are not consistent. Leite and Cavalheiro (1995) and Chakir et al. (2006) reported that CBZ reduced the frequency of spontaneous seizures in epileptic rats, but only one dose (40 mg/kg intraperitoneal three times/day; i.e., 120 mg/kg/day) was tested. In contrast, Nissinen and Pitkanen (2007) found no effect of CBZ (120 mg/kg/day, gavage) on spontaneous seizure frequency, and only 29% (2 of 7) of the rats were responders (i.e., animals exhibiting a reduction of ≥50% in seizure frequency), whereas the remaining animals (5/7) had an increase in seizure frequency. Some animals in their study had a low baseline seizure frequency, and a large difference in baseline seizure frequency across animals could have increased variance and thus reduced resolution. In Grabenstatter et al. (2007), single doses of CBZ administered in food were as effective as intraperitoneal injections at suppressing convulsive seizures. In the present study, CBZ (30 mg/kg/day) reduced the frequency of convulsive seizures to about 50%; 100 mg/kg/day nearly blocked all convulsive seizures, and a high dose of CBZ (300 mg/kg/day) did block all convulsive seizures. Anecdotal observations suggested that some of the animals treated with 300 mg/kg/day CBZ may have been sedated during the CBZ treatment, but no obvious behavioral changes were observed at lower doses (30 and 100 mg/kg/day). The issue of possible sedation, however, was not explored further. The higher control values for the trials with CBZ at 30 and 100 mg/kg/day versus 300 mg/kg/day probably arose because two different animals with higher baseline seizure frequencies were used. The single daily feeding of 300 mg/kg/day of CBZ was as effective as the 100 mg/kg/day three times per day treatment regimen used by Grabenstatter et al. (2007). We aimed to eliminate a dose in the middle of the night by providing all of the food (i.e., the entire dose) in the morning (9:00 a.m.). Comparison of these data suggests that CBZ administration in a once-per-day protocol yields a comparable anticonvulsant effect to three times per day CBZ treatment.

Plasma CBZ levels and correlation to sustained anticonvulsant effect

The therapeutic range of CBZ in epileptic patients has been estimated to be 4–10 μg/ml (Hvidberg, 1985; Honack & Loscher, 1989). In the present study, we established that the drug-in-food method of AED administration resulted in a dose-dependent correlation with plasma CBZ levels. Because rats normally consume more food at night, plasma levels of drug would otherwise be expected to drop during the day (Loscher, 2007); however, with the present feeding approach, near-therapeutic levels were probably sustained throughout the 24-h day. Human therapeutic levels measured 8 h after CBZ food administration were achieved with only the 300 mg/kg/day treatment, but the anticonvulsant effect of CBZ treatment was maintained for 24 h with all of the doses tested except 30 mg/kg/day, which did not protect early in the morning (6–10 a.m.). The animals had probably eaten all of the food earlier, resulting in a gap in CBZ ingestion and low plasma levels, and thus a reduced effect on convulsive seizures. The plasma levels of CBZ (30 and 100 mg/kg/day) at 5 p.m. (i.e., before the peak level of eating at 7–9 pm) might have been low if the animals ate relatively little of the drug-containing food during the day (i.e., light hours). Our findings of significant seizure suppression despite the relatively low plasma CBZ concentrations, however, are consistent with an earlier report where CBZ reduced hexafluorodiethyl ether–induced seizures, even though CBZ levels measured in plasma and brain were below the human therapeutic range (i.e., <2 μg/ml, Carl & Smith, 1989). Grabenstatter et al. (2007) also unexpectedly observed a persistent effect of a single administration of CBZ (100 mg/kg) for 20 h, despite the short elimination half-life of CBZ. They did not, however, determine the plasma concentrations after the CBZ-in-food administration. In the present study, plasma CBZ levels were estimated only at one time point, since the objective was to ascertain whether CBZ administered in food results in dose-dependent plasma concentrations. However, it would be useful to determine the plasma levels achieved at different time intervals across the 24-h day in epileptic rats to better explain the sustained anticonvulsant effect observed in the present work.

Apparent rebound effect after CBZ withdrawal

Grabenstatter et al. (2007) reported an apparent overshoot in seizure frequency after a single dose of CBZ. Likewise, Honack and Loscher (1989) found a significant increase in the duration of seizures and afterdischarges following withdrawal of CBZ treatment in kindled rats, but the degree to which this can be attributed to an actual “overshoot” in seizure frequency or probability is unclear. We observed an apparent increase in seizure frequency after an abrupt withdrawal from 5 days of CBZ (30 mg/kg/day) treatment. This apparent increase was probably not an actual rebound effect, however, because the raw data indicated that the control values used for comparison were lower than the baseline values. No apparent overshoot was observed at higher doses (100 or 300 mg/kg/day), and the seizure frequency appeared to return gradually to control values during the two recovery days (i.e., postdrug days). In order to determine whether an actual rebound effect occurs after CBZ treatment, a longer recovery time would be required. The lower postdrug seizure frequency (with 100 and 300 mg/kg/day treatment) than predrug values may be due to the persistent presence of CBZ in the brain even after its withdrawal.

Conclusions and future studies

The once-per-day AED-in-food protocol should be effective for both short- and long-term drug administrations in epileptic rats. This method of AED administration minimizes stress, which is associated with handling for intraperitoneal or gavage administration, and is potentially useful in chronic AED testing. CBZ plasma levels were measured at one time point only, but dose–dependent concentrations could be achieved with this protocol. Additional studies are warranted to determine whether therapeutic concentrations are maintained over a 24-h period. Although no overt behavioral toxicity was observed at lower doses of CBZ (30 and 100 mg/kg/day), toxicity tests should be conducted to ascertain the clinical relevance of these doses. Although the present study showed the utility of behavioral monitoring to study convulsive seizures, and CBZ clearly suppressed spontaneous convulsive seizures, the effect of long-term treatment on nonconvulsive electrographic seizures with this protocol remains to be determined. Nonetheless, these experiments provide a novel method for chronic, long-term administration of AEDs that is simple, inexpensive, reliable, and noninvasive.


This work was supported by grants from the National Institutes of Health (NS049620 and NS045144). Infrastructure support was also from NINDS contract NO1-NS-4-2359. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.


FED has received support from Johnson Pharmaceutical Research Institute, Johnson-Ethicon, and Neurotherapeutics Pharma; has been a paid consultant from Epitel, Inc. and Neurotherapeutics Pharma; and, has equity interest in Epitel, Inc. The remaining authors have no conflicts of interest.